* common.mk: add a rule for explicit.o.
* configure.in: detect explicit_bzero and memset_s.
* include/ruby/missing.h: add explicit_bzero.
* random.c (init_randomseed): use explicit_bzero() instead of
memset(). memset could be eliminated by compiler optimization.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52806 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk: update dependency of string.o to id.h for r52479.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52482 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (incs): remove extra slash before
vm_call_iseq_optimized.inc. as nmake vpath syntax includes a
separator, don't append it.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52262 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
vm_call_iseq_setup_normal_0start() is simple, however it has
some loops/conditions depends on ISeq::param.size and
ISeq::local_size (in vm_push_frame(), inlined into this function).
There are many simple methods which has a few parameters and local
variables. So that this patch introduces several special functions
generated in vm_call_iseq_optimized.inc by
tool/mk_call_iseq_optimized.rb.
This script makes
vm_call_iseq_setup_normal_0start_Xparams_Ylocals()
where X is 0 to 3 and Y is 1 to 6 (as current setting).
In this case, X * Y = 24 functions are created.
These functions creates fast method dispatch by inlining
vm_push_frame() with immediate params/locals sizes.
On my laptop, we can have the following results.
vm2_method* 1.083 (8.3% faster)
vm2_poly_method* 0.961 (3.4% slower)
It shows 8.3% faster for inner loop method dispatch (hit inline
cache), but 3.4% slower when inline cache miss because we need
to find a suitable call handler.
* common.mk: add a rule for vm_call_iseq_optimized.inc.
* tool/mk_call_iseq_optimized.rb: added.
* vm.c: include vm_call_iseq_optimized.inc.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52254 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (showflags): do not show flags from recursive make.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52221 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Even if S_ISREG() is true, the file may be file on FUSE filesystem
or something. We can't assume O_NONBLOCK is safe.
Moreover, we should wait if the path is point to FIFO. That's
FIFO semantics. GVL should be transparent from ruby script.
Thus, just reopen without O_NONBLOCK for filling the requirements.
[Bug #11060][Bug #11559]
* ruby.c (loadopen_func): new for the above.
* file.c (ruby_is_fd_loadable): new. for checks loadable file type
of not.
* file.c (rb_file_load_ok): use ruby_is_fd_loadble()
* internal.h: add ruby_is_fd_loadble()
* common.mk: now, ruby.o depend on thread.h.
* test/ruby/test_require.rb
(TestRequire#test_loading_fifo_threading_success): new test.
This test successful case that loading from FIFO.
* test/ruby/test_require.rb
(TestRequire#test_loading_fifo_threading_raise): rename from
test_loading_fifo_threading. You souldn't rescue an exception
if you test raise or not.
Moreover, this case should be caught IOError because load(FIFO)
should be blocked until given any input.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52151 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* file.c (rb_file_load_ok): open in non-blocking mode withoout
releasing GVL. don't care about others than regular files and
directories. [ruby-dev:49272] [Bug #11559]
* ruby.c (load_file_internal): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52139 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (check): separate test-testframework from test-all
only when building check.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52096 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (test-almost): rename target to test other than
test-frameworks.
* defs/gmake.mk (test-almost): run after test-testframework.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52022 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (check): separate test-testframework and test-all.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52006 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Version for character normalization data from 7.0.0 to 8.0.0.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52000 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk: clear environment variables for GEM paths to get rid
of searching unexpected gems. fix failures on travis-ci.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51946 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* enc/{ascii,us_ascii,utf_8}.c: set encoding indexes of
fundamental built-in encodings so that usable as well as
allocated rb_encoding before rb_enc_init().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51862 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* encindex.h: separate encoding index constants from internal.h.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51861 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (yes-test-ruby): apply excluding options as well as
"test".
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51855 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_w32_reparse): read reparse point in a dynamic
buffer.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51676 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
and ConditionalVariables to thread_tools.c. In other wowrds,
such classes are built-in.
[Feature #8919]
At first, I planned to embed only a Queue class.
However, rubygems requires 'thread.rb' (rubygems are
required at first, when launch MRI without --disable-gems).
So most of people require 'thread.rb' as an embedded library.
Now, ext/thread/thread.c is empty, only for a dummy for
compatibility.
* thread.c: move a definition of Mutex class to thread_tools.c.
And define Mutex class under Thread (so now Mutex is Thread::Mutex).
Because other thread related classes are also defined under Thread.
We remain ::Mutex as Thread::Mutex. Only an inspect result is changed.
* common.mk: add dependency from thread.o to thraed_tools.c.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51660 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
[Feature #11420]
This table only manage ID->VALUE table to reduce overhead of st.
Some functions prefixed rb_id_table_* are provided.
* id_table.c: implement rb_id_table_*.
There are several algorithms to implement it.
Now, there are roughly 4 types:
* st
* array
* hash (implemented by Yura Sokolov)
* mix of array and hash
The macro ID_TABLE_IMPL can choose implementation.
You can see detailes about them at the head of id_table.c.
At the default, I choose 34 (mix of list and hash).
This is not final decision.
Please report your suitable parameters or
your data structure.
* symbol.c: introduce rb_id_serial_t and rb_id_to_serial()
to represent ID by serial number.
* internal.h: use id_table for method tables.
* class.c, gc.c, marshal.c, vm.c, vm_method.c: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51541 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (getifaddrs): use ruby_strdup instead of
combination ruby_xmalloc, lstrlen and lstrcpy.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51517 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (PHONY): should not depend on any real files, not to
make those dependencies inadvertently.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51447 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk ($(arch)-fake.rb): read from STDIN instead of creating
version.i, to get rid of conflict with tool/update-deps.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51419 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This speeds up the hash function for dynamic symbols.
[ruby-core:70129] [Bug #11396], nearly up to Ruby 2.1 levels
Power-of-two hash sizing [Feature #9425] speeds up cases where we
have a good hash, but this means we can no longer hide behind weak
hashes. Unfortunately, object IDs do not hash well, but we may
use the extra space in the RSymbol struct to memoize the hash value.
Further optimizations should be possible. For now, the st.c APIs
force us to calculate rb_str_hash redundantly at dsym registration.
* symbol.h (struct RSymbol): add hashval field
* symbol.c (dsymbol_alloc): setup hashval field once
* hash.c (rb_any_hash): return RSymbol->hashval directly
* common.mk: hash.o depends on symbol.h
Thanks to Bruno Escherl <bruno@escherl.net> for the bug report
[ruby-core:70129] [Bug #11396]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51410 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (distclean-local): clean preludes in the build
diretory, not by realclean.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51400 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (clean, distclean): clean local files before platform
directory, some files may exist there.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51399 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk: fix macro name, exec but not EXEC. pointed by
@nagachika.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51336 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* random.c (fill_random_bytes_syscall): try getrandom system call
on Linux if supported by the kernel.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51182 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* common.mk (ENC_MK): reorder $(ENCS) before the target file name,
as enc/make_encmake.rb ignores any arguments after the target.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50954 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
kosaki
usa
hsbt
nobu
ko1
kazu
duerst
svn
normal